1. Technical Field
The present disclosure relates to a method for coupling a graphene layer and a substrate, and a device comprising said graphene/substrate structure.
2. Description of the Related Art
In the last few years, application of surface-treatment techniques has made it possible to improve the properties of materials, involving an extremely wide range of industrial sectors, from the mechanical sector to the biomedical sector. The development of such treatments is the result of a considerable progress in research, pushed by the desire to increase the performance and potentialities of the surfaces treated and by the desire to reduce costs.
Application of such treatments, associated to the study of the interfaces of the materials involved, is fundamental in the development of systems that present a high surface/volume ratio since the phenomena that occur at the interface affect the physics of the system under examination, all the more, the greater said ratio.
Among the materials that have proven of particular interest, graphene, a material constituted by a single layer of carbon atoms arranged in a hexagonal crystalline lattice, presents surprising electrical, optical, and mechanical properties due to its two-dimensional structure.
Graphene can be obtained via processes of mechanical exfoliation, graphitization of silicon carbide (SiC), growth on metal substrates by means of chemical vapor deposition (CVD), exfoliation in liquid phase, and opening of carbon nanotubes.
Where the technique of production of graphene is based upon processes of CVD growth on metal substrates, or exfoliation techniques, in addition to the problem of choosing an adequate technique of transfer of the graphene onto a substrate, there is posed also the problem of identifying surface treatments capable of promoting adhesion of the graphene on the desired substrate, and at the same time, modulating appropriately the phenomena that occur at the graphene/insulator interface so that the electrical properties of the graphene will not be impaired by the treatment adopted.
Typically, a surface treatment is based upon techniques that aim at modifying the composition of a surface, incorporating elements or functional groups that modify the properties of the surface, but not those of the material as a whole.
In many cases, the surface treatment has the specific purpose of promoting adhesion between surfaces that are dissimilar so as to guarantee the temporary or permanent bonding between them.
For the adhesion between two surfaces to be effective, the surfaces must be clean, smooth, and chemically receptive. This is obtained through the aid of a wide range of techniques that may be of a mechanical, chemical, or physical type.
The mechanical methods, generally based upon abrasion of the surface, are very rough and invasive. These approaches, albeit modifying in a macroscopic way the surface (operating on the roughness, thickness, or planarity) frequently generate debris that abundantly contaminates the surface itself.
The procedures that presuppose the aid of chemical solutions use techniques of dipping, spin-coating, or casting of the surface to be treated in/with an appropriate solution.
Techniques of “self-assembly” of molecules by adopting the L-B (Langmuir-Blodget) procedure also form part of this strategy.
One of the physical treatments that are most widely used in the industrial field for deposition of functional coatings is that of vacuum plasma treatments.
The so-called “plasma” technology envisages the use of a partially ionized gas (plasma), made up of a mixture of electrons, positively and negatively charged particles, neutral atoms, and molecules. Each of said components can trigger chemical and physical reactions on the surface of the materials with which it comes into contact, and, in given conditions, can generate transient or permanent modifications.
This technique proves more versatile due to the possibility of modulating the surface properties (hydrophilia/hydrophobia, density of the functional groups, and stability of the film) by changing the process parameters (for example, power and modulation of the discharge, monomer/process gas ratio).
The selection of the appropriate procedure is made on the basis of the nature of the substrates involved and the type of interface that is to be obtained.
Since, in the case of graphene, it is a monolayer of carbon atoms, its electronic properties are strictly correlated to the nature of the substrate, to the gases, or more in general to the molecular environment that surrounds it.
Silicon oxide (SiO2), the substrate that is commonly chosen for transfer and integration of the graphene, provides an excellent optical contrast, which, however, albeit enabling visual identification of the presence of the graphene, presents disadvantages linked to the surface roughness, to the high surface-charge concentration, to the presence of phonons, to the hydrophilic properties of the material, and to the low dielectric constant (κSiO2=3.9), which jeopardize the electrical performance thereof.
Effects of scattering due to the surface charge of the silicon oxide are in part responsible for the reduction of mobility of the carriers of the transferred graphene as compared to the carriers of the suspended graphene (2.105 cm2/V·s).
Experiments have moreover revealed the key role of the morphological structure of the silicon oxide: STM (Scanning Tunneling Microscopy) measurements suggest in fact that the graphene follows the roughness of the silicon oxide. Moreover, experiments on electro-mechanical systems made of graphene indicate that the substrate induces significant stresses in the graphene itself.
Another phenomenon that affects the electrical properties of the graphene transferred onto silicon oxide is the capacity of this for trapping chemical species. The surface of the oxide terminates, in fact, with hydroxyl polar groups, OH—, so that it attracts polar molecules like water. The surface of the oxide hence has one or more layers of water absorbed on the interface that is formed between the silicon oxide and the graphene.
The graphene/silicon oxide interface is undoubtedly complex not only on account of the chemico-physical properties of silicon oxide, but also on account of the hydrophobic behavior of the graphene itself (the angle of contact measured on the graphene obtained by mechanical cleavage is 91°). The poor adhesion of graphene on silicon oxide thus imposes a pre-treatment of the oxide for promoting adhesion and at the same time minimizing the chemico-physical interactions that jeopardize the electrical properties of the graphene.
In the past, different strategies have been proposed for favoring adhesion of the graphene to silicon-oxide surfaces, the most common of which (applied also on Al2O3 and GaN), has the purpose of cleaning the surface through a wet treatment followed by an oxygen plasma. The procedure envisages ultrasound cleaning (referred to as “SC-1 treatment”), followed by oxygen-plasma treatment.
The SC-1 treatment consists of cleaning with a mixture of water (H2O), hydrogen peroxide (H2O2), and ammonium peroxide (NH4OH).
The mixture is typically constituted by 5:1:1 parts per volume of H2O, H2O2, and NH4OH. The treatment envisages dipping of the substrate for 5 to 10 minutes in the solution heated to 70° C.-75° C., followed by rinsing in deionized water. The SC-1 treatment eliminates from the surface of the oxide organic contaminants through the solvatating effect of ammonium peroxide and the oxidizing action of hydrogen peroxide. The ammonium peroxide also has the task of removing metal elements, such as Cu, Au, Ag, Zn, and Cd, Ni, Co, Cr. The subsequent oxygen-plasma treatment has the purpose of cleaning the silicon-oxide surface and of modifying the chemical and electrostatic characteristics thereof. The effects that are noted on the treated surface are removal of the organic contaminants and the residual layers, increase of the surface tension, through reduction of the angle of contact in regard to liquids, and formation of a surface capable of reacting actively with the polymers.
Another approach presupposes the introduction of an additional treatment under hexamethyldisilazane (HMDS), which is carried out following upon the standard treatment that envisages SC-1 and oxygen plasma.
This procedure, by means of introduction of methyl groups, renders the silicon-oxide surface more hydrophobic, reducing the absorption of water at the interface formed by the graphene and the oxide.
In general, the most common wet treatment (SC-1) followed by oxygen plasma presents major limits of applicability. Typically, it is used on a non-structured substrate of SiO2 and without metal contacts or interconnections, since the presence of solutions, such as H2O2 and the oxygen plasma itself, may dramatically jeopardize the pre-existing structure both from a morphological standpoint (roughness, wettability, phenomena of corrosion, chemical etching of the exposed surfaces, oxidation phenomena) and from an electrical standpoint (integrity of the metal surfaces, contact resistance).
In particular, the SC-1 treatment etches the following metals: Al, Cu, Ag, Ti. The following metals are not, instead, sensitive to this chemical treatment: Au, Pt, and Ni.
Consequently, in the devices obtained with the aid of graphene, whatever the application, there emerges a desire for architectures on SiO2, which envisage integration of the graphene on the oxide as single and non-structured interface. To overcome the limits linked to poor adhesion of graphene on SiO2, thus preventing introduction of pre-treatment of the surfaces, architectures are frequently favored that presuppose integration of suspended graphene between the source and drain contacts (for example, in production of a FET).
In this case, also the mobility of the carriers is indisputable favored, due to the absence of problematical interfaces, such as the SiO2. However, this is an approach that cannot be applied over extensive areas (the size of the suspended graphene layer is tiny, just a few microns), it being markedly dependent upon the geometrical dimensions of the device.
Another approach is the one proposed in EP2362459, also published in U.S. Patent Publication No. 2012/058350, which makes it possible to overcome the limits linked to the interaction of the graphene with the surrounding environment (for example, absorption of oxygen and water) and at the same time promote its adhesion on substrates such as SiO2. It is based upon a process of non-covalent functionalization of the silicon-oxide substrate on which the graphene with organic molecules is to be transferred. Said organic molecules are provided with groups that bind to the silicon oxide, such as for example diaminodecane.
In particular, formed on the silicon-oxide surface is a monolayer of molecules comprising an anchorage group, which creates a non-covalent bond with the graphene, a functional group, and an alkyl spacer group, which unites the functional group and the anchorage group and facilitates formation of the monolayer, promoting stability thereof.
Formation of this layer is basically due to the interactions of the amino groups with the surface of the substrate of SiO2 via the Van der Waals forces. The monolayer may be formed by dipping the substrate in a solution containing the molecules to be immobilized or by depositing said solution, which is then left to act for a preset period of time.
A coating layer is thus created, having a dense and very stable structure, over the entire surface of the substrate.
In detail, the procedure of surface treatment illustrated in EP2362459 comprises the following steps:
dipping the substrate of oxidized silicon in a 10-mM solution of 1,10-diaminodecane for a duration of three hours;
rinsing in a mixture of solvents (1:10 methanol tetrahydrofuran), which has the purpose of removing the excess 1,10-diaminodecane; and
drying in nitrogen gas.
This procedure presents the major limit of using dangerous chemical products, such as for example tetrahydrofuran, which reacts violently with oxidizing agents, generating explosions. It is moreover highly flammable, and may give rise to phenomena of uncontrolled polymerization.
The National Toxicology Program (NTP) of the U.S. Department of Health and Human Services has moreover highlighted that, in studies of inhalatory exposure of rats (two years), there has been observed a carcinogenic activity.
Another approach suggested by Liang et al. (Nanoletters, 2007) presupposes the introduction of low-viscosity thermal glues with a typical thickness of 10 nm.
After the process of thermal stabilization, the glue laid by casting or spin-coating on the surface of SiO2 becomes solid and functions as dielectric in the graphene/silicon oxide interface.
This approach, which is rarely applied, presents the disadvantages typical of polymeric materials, namely, low temperature of vitreous transition, which hinders implementation of high-temperature processes, non-applicability on accentuated topography (>>10 nm), creation of a dielectric of lower quality than the insulators (gate oxide) most commonly used for producing an electronic device.
It is consequently felt in the art a desire to find a method alternative to those so far used for favoring coupling of a graphene monolayer and a substrate that is free from the disadvantages described above. In particular, there is felt the desire for a method that is simple and inexpensive, does not involve the use of toxic solvents, and may be used for deposition of graphene over extensive areas of a substrate, without altering the pre-existing structures either from the morphological standpoint or from the electrical standpoint.